Radtask Mcp

io.github.dougsureel-techPublished by GitHub account dougsureel-techv0.1.0Updated Oct 24, 2026

AI task cockpit: an agent does your tasks and asks before anything irreversible.

VerifiedStreamable HTTPWeb executableProductivity & WorkflowAI & ML

Overview

AI-generated overview

Remote MCP server for RadTask, an AI task cockpit where an assistant can create, list and complete tasks that a separate agent then carries out.

What you can do
  • Create tasks in the user's RadTask cockpit from plain-English descriptions, with an optional project.
  • List the user's tasks, optionally filtered by status.
  • Mark a task done by its id, taken from the task list.
  • Tasks land in a cockpit where a separate RadTask agent researches, drafts, writes and schedules, holding irreversible steps for user approval.
When to use it
Useful when you want an assistant to capture and track to-dos in a dedicated task manager rather than a chat, and to hand the actual work to an agent that asks before sending, spending, deleting or publishing.
Requirements
A remote Streamable HTTP endpoint at radtask.ai; no local package. A private token obtained from the provider's connect page, passed as a query parameter in the endpoint URL. The service is described as a private beta.
Before you install
The token in the endpoint URL acts as the cockpit's namespace, so anyone with it can add tasks to that account; keep it out of shared configs and logs. The server only creates, lists and completes tasks, but the linked agent may take real actions such as sending, spending, deleting or publishing, which the provider says are held for human approval. The provider states this is a safety design, not a guarantee, and the service is in private beta.

Installation

In SourceWeft

  1. Open Radtask Mcp in the dashboard and add it to a workspace.
  2. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/io.github.dougsureel-tech%2Fradtask-mcp/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config.

{
  "mcpServers": {
    "radtask": {
      "type": "http",
      "url": "https://radtask.ai/api/mcp"
    }
  }
}

README

RadTask MCP Server

Connect any AI to your task cockpit. RadTask is an AI task manager whose agent actually does the task — and asks first before anything irreversible. It runs a live MCP (Model Context Protocol) server, so Claude, ChatGPT, Cursor, or your own agent can create and track tasks that land in the user's RadTask cockpit, where a safe autonomous agent runs them.

Private beta. A tool with a safety design — not a guarantee.

Connect

Endpoint (Streamable HTTP, JSON-RPC 2.0):

https://radtask.ai/api/mcp?token=YOUR_TOKEN

Get your private token at radtask.ai/connect. The token is your cockpit's namespace — tasks an AI adds appear in your RadTask.

Claude Desktop / Cursor

json
{  "mcpServers": {    "radtask": { "type": "http", "url": "https://radtask.ai/api/mcp?token=YOUR_TOKEN" }  }}

Tools

ToolWhat it does
create_taskCapture a task in plain English (optional project)
list_tasksList tasks, optionally filtered by status
complete_taskMark a task done by id

Why RadTask is different

Most "task manager" MCP servers are queues — they track what an AI should do. RadTask does the task: a real agent researches, drafts, writes documents, and schedules in a sandbox — and a deterministic risk classifier holds only the irreversible steps (sending, spending, deleting, publishing) for the user's one-tap approval.

The safety posture is published and machine-verifiable:

Design goal: zero unsupervised destructive actions. Irreversible actions are always held for a human.

Links

License

MIT © Sureel

Source: README.md at commit 27fd2f9

Tools

3

create_task

Add a task to the user's RadTask cockpit. Use this to capture anything the user wants done or tracked.

Parameterstitlerequiredproject

list_tasks

List the user's RadTask tasks, optionally filtered by status.

Parametersstatus

complete_task

Mark a RadTask task done by its id (from list_tasks).

Parametersidrequired

Version history

1
  1. v0.1.0LatestSep 16, 2026

Score

100% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listedRequired15 pts
  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts